Meta’s Muse AI assistant, which allows users to cancel unwanted subscriptions, has been downloaded over 3 million times. However, it raises significant security concerns. Users must provide Muse access to sensitive information like bank accounts, credit cards, emails, and texts. This requirement draws attention to the need for strong cybersecurity measures.
The appeal of AI is evident in its ability to handle tedious phone calls, such as resolving billing issues with a wireless carrier or making restaurant reservations online. Muse and Instinct, a similar AI service, aim to further this capability by making calls on your behalf. This can indeed ease tasks requiring multiple steps or prolonged waiting times.
A core concern is how comfortable users feel letting software speak for them and make decisions on their behalf.
Our CyberGuy LIVE classes have explored five ways AI can help in healthcare, enabling you to prepare for appointments, track details, understand medical information, research prescriptions, and organize follow-ups. You can watch these replays and more on cyber security tips at CyberGuyLive.com.
Instinct and Meta’s AI Features
Instinct Concierge, launched on September 16, allows its AI to handle phone calls and tasks such as restaurant reservations and managing accounts. The company received substantial investment, raising $250 million in a Series B round and $1 billion in a Series C, valuing it at $10 billion. Meta Muse, launched on September 8, can also browse the web and make calls for users.
Meta Muse’s popularity is indicated by its rapid ascent on Apple’s U.S. App Store, with over 3 million downloads. However, privacy concerns need addressing, given the sensitive data access these assistants require.
Cost and Practical Usage
Instinct has not disclosed its subscription price publicly, while Meta Muse offers a free tier with a limit, after which users may opt for paid plans. These services aim to provide convenience, but at the cost of granting substantial access to user information.
Why AI Calls are Appealing but Risky
AI can manage tasks without distracting users, but each additional function lets AI access more personal details. Users must be cautious about the extent of access granted, ensuring explicit limits.
Privacy and Security Concerns
User privacy policies allow AI to access connected app data. This includes messages and emails, which could expose comprehensive personal information. Instinct offers an opt-out option for data use in AI training, though previously trained models retain their improvements. Google Workspace data receives special treatment, excluded from AI model training.
Meta Muse runs in a secure cloud environment, with credentials secured. Muse’s actions are logged and reviewed, adding a layer of user control.
Considerations Before Granting AI Access
- Begin with low-risk calls to test AI performance.
- Review AI access to personal services and revise permissions regularly.
- Limit sensitive information exposure, especially financial data.
- Use confirmation controls for critical actions like purchases.
- Verify AI-initiated tasks post-completion for reliability.
- Be aware of data deletion differences when disconnecting services.
- Exercise caution with health and sensitive transactions.
- Keep updated on privacy settings with evolving AI features.
- Secure connected accounts with strong passwords and two-factor authentication.
- Install robust antivirus software as AI may interact online.
AI offers valuable time-saving benefits, but privacy and security risks require careful consideration. Users should start small, monitoring AI actions to ensure privacy and access control.
Leave a Reply